ON THE CHOICE OF THE MAIN DIMENSIONS OF AN IMPULSE GENERATOR

G. A. SIPAILOV, K. A. KHOR'KOV

(Presented by the scientific seminar of the Department of Electrical Machines and General Electrical Engineering)

Impulse-power generators intended for obtaining large powers in a single pulse, as well as breaking-power generators used for testing high-voltage oil circuit breakers, are manufactured on the basis of existing turbogenerators. The problems of designing turbogenerators have been developed quite fully, and this substantially facilitates the task of creating breaking- and impulse-power generators. However, the individual features determined by the operating conditions of breaking- and impulse-power generators must, naturally, be taken into account in their design; in particular, this applies to the choice of their main dimensions.

If for turbogenerators there exists a logical and mathematical justification for choosing the best computational and design-technological variants of the machine in its design [1], then for impulse- and breaking-power generators this problem requires further investigation and solution.

For example, for impulse-power generators, when choosing the optimal values of the diameter and length of the active part, the number and dimensions of the slots, etc., in a machine of specified overall dimensions, it is necessary to proceed from the condition of obtaining the maximum electromagnetic energy during the first half-period of the current. In this case, economic considerations (production cost and cost of losses), which substantially influence the choice of the main dimensions of series turbogenerators, play a secondary role, in view of the highly individual character of the machine. The thermal regime of the machine also imposes no special restrictions if the object on which the generator operates does not require frequent repetition of pulses. This makes it possible to exclude the influence of economic factors and heating questions on the relationship among the main dimensions of an impulse generator.

In the present work, results are presented for determining the optimal dimensions of an impulse-power generator, using the Minsk-1 computing machine, from the condition of obtaining maximum impulse power within specified overall dimensions. In deriving the equations on the basis of which the calculation program was compiled, the formulas of I. M. Postnikov [2, 3] and equations relating the power, energy, and machine constant of an impulse generator [5] were used.

Expression of the Inductive Reactance in Terms of the Geometric Parameters of an Impulse Generator

On the basis of the equations obtained in [5], the quantity proportional to the volume of the active part can be expressed in terms of the energy delivered by the generator to the load:

\[ D^{2}l=\frac{1}{n}P_{\mathrm{ud}}C_{\mathrm{ud}} =\frac{2\pi}{k_{\mathrm{n}}}Q_{\mathrm{n}}C_{\mathrm{ud}}, \tag{1} \]

where
\(Q_{\mathrm{n}}\) is the energy delivered to the load;
\(k_{\mathrm{n}}\) is the load factor;
\(C_{\mathrm{ud}}\) is the machine constant of the impulse generator, and for a single-phase generator

\[ C_{\mathrm{ud}}=\frac{C_A}{4k_3k_i k_m} =\frac{0.35}{k_3}x_{\mathrm{ud}}C_A, \tag{2} \]

\(C_A\) is Arnold’s machine constant;
\(k_3\) is the damping coefficient of the impulse current;
\(x_{\mathrm{ud}}\) is the inductive reactance of the impulse generator.

As follows from equations (1) and (2), for a constant machine volume the energy delivered to the load will be the greater, the smaller the value of the reactance \(x_{\mathrm{ud}}\).

As the initial equation in determining the inductive reactance of a single-phase impulse generator, one may take

\[ x_{\mathrm{ud}}=\frac{x'_d+x_2}{2}, \tag{3} \]

where \(x'_d\) is the transient inductive reactance along the longitudinal axis;
\(x_2\) is the negative-sequence inductive reactance.

In the analytical determination of the negative-sequence inductive reactance, it is usually expressed through the subtransient reactances along the longitudinal and transverse axes. The numerical values of the reactance \(x_2\) for turbogenerators, depending on the damping capacity of the rotor, lie within the limits [3, 4]

\[ x_2=(1 \div 1.24)x''_d. \]

Taking into account that an impulse generator has an enhanced damping system, the negative-sequence reactance in magnitude will approach the subtransient reactance along the longitudinal axis,

\[ x_2=(1 \div 1.1)x''_d. \tag{4} \]

In turn, the subtransient reactance is equal to

\[ x''_d=x_s+x''_{ad}, \]

where \(x_s\) is the leakage inductive reactance of the stator winding;
\(x''_{ad}\) is the subtransient inductive reactance of armature reaction.

The leakage reactance of the stator winding in turbogenerators amounts to about 90% of the value of the subtransient reactance along the longitudinal axis, that is,

\[ x''_d=1.1x_s. \]

Then, taking equality (4) into account, we obtain

\[ x_{\mathrm{ud}}=(1.1 \div 1.15)x_s. \tag{5} \]

To determine the leakage reactance of the stator winding, several methods exist; however, the following is usually taken as the starting equation [3]:

\[ X_s=4\pi f\mu_0\,\frac{1}{p}\,l\,\frac{w_1^2}{q_1}\,(\lambda_{\mathrm{p}}+\lambda_{\mathrm{l}}+\lambda_{\delta})\ \Omega, \tag{6} \]

where
\(w_1\) — number of series-connected turns of a phase;
\(\mu_0\) — magnetic permeability of the medium;
\(q_1\) — number of slots per pole and phase filled with copper;
\(p\) — number of pole pairs;
\(\lambda_{\mathrm{p}}, \lambda_{\mathrm{l}}, \lambda_{\delta}\) — leakage permeances, respectively slot, end, and differential.

For a single-phase single-layer winding

\[ w_1=pq_1 \quad \text{and} \quad q_1=\frac{1}{3p}z_1, \]

where \(z_1'\) is the total number of stator slots.

It should be noted that for the single-phase machine under consideration the number of slots per phase will be equal to the number of slots for two phases of a three-phase machine; that is, the single-phase winding will correspond to two phases of a three-phase winding. Consequently, the short circuit of such a single-phase machine will be analogous to a two-phase short circuit of a three-phase machine. This, in particular, explains the form of equation (3).

For convenience in programming the calculation on a computing machine, let us express the leakage permeances in a somewhat different form than usual.

The slot-leakage permeance for an open slot is

\[ \lambda_{\mathrm{p}}=\frac{1}{3}(k_1+2k_2+k_{\mathrm{i}})k_{\mathrm{ns}}, \tag{7} \]

where
\(k_1\) — ratio of the slot height to its width;
\(k_2\) — ratio of the wedge height to the slot width;
\(k_{\mathrm{i}}\) — ratio of twice the insulation thickness to the slot width;
\(k_{\mathrm{ns}}\) — coefficient taking into account saturation of the steel of the stator tooth layer.

The end-leakage permeance for a single-layer subdivided concentric winding is

\[ \lambda_{\mathrm{l}}= \left( 0.13-\frac{0.1}{p}+0.039\,k_{\mathrm{p}}+0.33\,\frac{k_1k_{\mathrm{p}}}{z_1'} \right)p\beta^2 z_1' k_g, \tag{8} \]

where
\(k_{\mathrm{p}}\) — ratio of the slot width to the tooth pitch;
\(\beta\) — ratio of the pole pitch \(\tau\) to the length of the active part of the machine \(l\);
\(k_g\) — coefficient taking into account the damping action of the shielding rings, “caps,” on the leakage fields of the stator winding.

The differential-leakage permeance is

\[ \lambda_{\delta}=\frac{k^2 w_1 t_1}{\pi^2\delta k_{\delta}}(\lambda_z' + \lambda_{\mathrm{v}}), \]

where
\(k_{w1}\) is the winding factor of the fundamental harmonic;
\(t_1\) is the tooth pitch;
\(\delta\) is the air gap;
\(k_\delta\) is the air-gap coefficient;
\(\lambda_z\) is the leakage permeance for tooth harmonics, where

\[ \lambda_z=\left(\frac{z'_1}{2p}\right)^2 \sum_\nu \left(\frac{1}{\nu_z}\right)^2; \]

\(\nu_z\) is the order of the tooth harmonic;
\(\lambda_{\mathrm{v}}\) is the leakage permeance for space harmonics, where

\[ \lambda_{\mathrm{v}}=\left(\frac{z'_1}{2p}\right)^2 \sum_n^\kappa \left(\frac{k_{w\nu}}{k_{w1}}\cdot\frac{1}{\nu}\right)^2 k_{\mathrm{e}}. \]

The lower limit of the sum in determining \(\lambda_{\mathrm{v}}\) is \(n=5\); then all odd harmonics not multiples of three are added up to the value of the tooth frequency, with

\[ \frac{k_{w\nu}}{k_{w1}}= \frac{\sin \nu \frac{\pi}{3}\cdot \sin \frac{\pi}{2}Q} {\sin \frac{\pi}{3}\cdot \sin \nu \frac{\pi}{2}Q}, \]

where \(Q\) is the integer number of phases per pole;
\(k_{\mathrm{e}}\) is a coefficient taking into account the attenuation of higher harmonics damped by the damper system.

For impulse generators, as for turbogenerators, one may take \(k_{\mathrm{e}}\approx 0.6\); \(\lambda_z'=0.82\). For the value \(q/Q=2/3\) for a single-phase winding we obtain \(k_{w1}=0.825\); then the sum

\[ \varphi=\sum_{\nu=n}^{\kappa} \left(\frac{k_{w\nu}}{k_{w1}}\cdot\frac{1}{\nu}\right)^2 \]

will have the values presented in Table 1.

Table 1

Denoting the ratio \(\dfrac{\delta}{D}=k_i\) and expressing the tooth pitch through the diameter, for \(p=1\) we obtain

\[ \lambda_\delta= \frac{0.22}{k_i z'_1 k_\delta} \left(0.82+0.15\varphi z_1'^2\right). \tag{9} \]

Thus, the leakage reactance of the stator winding, expressed in terms of the geometrical parameters of the machine in the SI system of units, will be

\[ X_s=\frac{4.13\cdot 10^{-4}}{p^2\beta}\, z'_1\left(\lambda_{\mathrm{p}}+\lambda_{\mathrm{l}}+\lambda_\delta\right)D. \tag{10} \]

In relative units,

\[ x_s = X_s \frac{I_n}{U_n}, \]

where \(I_n\) is the nominal base current

\[ I_n = A t_1 = A \frac{\pi D}{z_1}; \]

\(U_n\) is the nominal voltage

\[ U_n = 39.7\, z_1 \tau e B_\delta, \]

\[ \frac{I_n}{U_n} = \frac{0.032\,p^2\beta}{z_1^2 D}\cdot \frac{A}{B_\delta}. \]

Then

\[ x_s = \frac{13.2\cdot 10^{-6}p}{z_1'} (\lambda_n+\lambda_l+\lambda_\delta)\frac{A}{B_\delta}. \tag{11} \]

The reactance of the impulse generator in relative units, in accordance with equations (5) and (11), will be

\[ x_{y,l} = \frac{15\cdot 10^{-6}p}{z_1'} (\lambda_n+\lambda_l+\lambda_\delta)\frac{A}{B_\delta}. \tag{12} \]

In the course of the calculation, for each machine volume we specify the corresponding values of the linear loading \(A\) and the induction in the air gap \(B_\delta\). For one machine volume (one rated power of the generator), the ratio of the linear loading to the induction will be constant, which is very important in analyzing the geometry of the machine.

Expression of the impulse power through the geometrical parameters of the generator

The power of the impulse generator can be expressed through the rated power of a three-phase synchronous machine [5]:

\[ P_{yд}=4k_3 k_i k_m P_3 = \frac{4k_3}{\sqrt{2}\,x_{yд}}\,P_3, \tag{13} \]

where \(P_3\) is the rated power of the three-phase machine [3]

\[ P_3=\frac{\pi^2}{\sqrt{2}}\,k_{w1}AB_\delta D^2 l n; \tag{14} \]

\(k_3\) is the current attenuation coefficient [5]

\[ k_3 = \frac{1+e^{-t/T_d''}+2e^{-t/T_a}}{4}. \tag{15} \]

The attenuation time constants \(T_a\) and \(T_d''\) of the aperiodic and subtransient components of the current can also be expressed through the geometrical dimensions of the machine. Then, for the instant of time \(t=\frac{\pi}{\omega}\), we obtain

\[ \frac{t}{T_a} = \frac{ 0.225\,z_1^2 \left( 0.89+\frac{1}{p\beta} +\frac{2.1k_l k_n}{z_1} +0.167k_n \right)\cdot 10^{-4}D }{ k_1 k_{\text{zap}} k_n^2(\lambda_n+\lambda_l+\lambda_\delta)\,D^2 l }; \]

\[ \frac{t}{T_d''}= \frac{\left[0.5+0.07\beta\,(z_1'-4)\right](z_1'+4)^2\cdot 10^{-4}} {\left[0.4+0.06\beta\,(z_1'+4)\right]\cdot D^2}. \]

Substituting into equation (13) the values of the quantities entering it from (12) and (14), we obtain a formula for determining the impulse power in volt-amperes

\[ P_{\text{ud}}= \frac{54.5\cdot 10^6 k_{\text{d}}D^2 l z_1' B_\delta^2} {p^2(\lambda_{\text{p}}+\lambda_{\text{l}}+\lambda_\delta)}, \tag{16} \]

where \(D\) and \(l\) are the stator bore diameter and the length of the active part, in meters;

\(B_\delta\) is the induction in the air gap, in teslas.

Determination of the optimal geometric ratios of an impulse generator

By the optimal geometric ratios of an impulse generator are meant such values of \(\beta\), \(k_1\), etc., at which the maximum power and, correspondingly, the maximum energy delivered to the load are achieved in a generator with a specified volume of the active part.

On the basis of the expressions obtained (7–16), an analysis was carried out of the influence of the slot geometry and the number of stator slots, the ratio between the principal dimensions, and other generator parameters on the value of the impulse power. The analysis was performed with the aid of the Minsk-1 digital computer. The block diagram of the calculation is presented in Fig. 1.

In the calculations the following limits of variation of the variables were adopted:

1. A quantity proportional to the volume of the machine, in \(m^3\).

\[ D^2 l = 1.0;\ 3.5;\ 7.0;\ 9.1;\ 11. \]

1. Outer diameter of the active steel of the stator, in m

\[ D_a = 1+k\cdot 0.2 \quad \text{to } D_a=2.6. \]

1. Total number of stator slots

\[ z_1' = 24+k\cdot 6 \quad \text{to } z_1'=72 \]

and, correspondingly, the total number of rotor slots

\[ z_2' = z_1'+4. \]

1. Ratio of the slot height to its width

\[ k_1=\frac{h_{\text{p}}}{b_{\text{p}}}=0.5+k\cdot 0.5 \quad \text{to } k_1=4. \]

1. Ratio of the pole pitch to the length of the active part of the machine

\[ \beta=\frac{\tau}{l}=0.2+k\cdot 0.1 \quad \text{to } \beta=0.9. \]

1. Damping coefficient

\[ k_{\text{d}}=0.5;\ 0.75;\ 1.0. \]

The coefficient \(k\) entering the expressions for \(D_a\), \(z_1'\), \(k_1\), and \(\beta\) is an integer taking successively the values

Fig. 1. Block diagram for calculating impact power on the Minsk-1 electronic digital computer

Flow sequence:

Left branch

Loading \(D^{2}\ell(0)\)
↓
Loading \(D_{a}(0)\)
↓
Loading \(Z_{i}'(0)\)
↓
Loading \(K_{1}(0)\)
↓
Calculation of \(D\)
↓
Calculation of \(\lambda_{p}\)
↓
Calculation of \(\lambda_{p}\)
↓
Calculation of \(\lambda_{0}\)
↓
Calculation of \(X_{s}\) (ohms)
↓
Calculation of \(\dfrac{I_{n}}{U_{n}}\)
↓
Calculation of \(X_{s}\) (p.u.)
→
Calculation of \(P_{\mathrm{gab.}}\)

Right branch

Calculation of \(P_{\mathrm{gab.}}\)
↑
Calculation of \(K_{\mathrm{zap.}}\)
↑
Calculation of \(P_{\mathrm{ud.}}\)
↑
Increment of \(K_{1}\)
↑
Comparison with \(K_{1,\max}\)
↑
Increment of \(Z_{i}\)
↑
Comparison with \(Z_{i,\max}\)
↑
Increment of \(D_{a}\)
↑
Comparison with \(D_{a,\max}\)
↑
Transition to a new \(D^{2}\ell\) by the counter
↑
Stop

Loop conditions shown in the diagram:

• from calculation of \(\lambda_{0}\) to comparison with \(K_{1,\max}\): if \(K_{1}<K_{1,\max}\);
• from comparison with \(Z_{i,\max}\) to loading \(K_{1}(0)\): if \(Z_{i}<Z_{i,\max}\);
• from comparison with \(D_{a,\max}\) to loading \(Z_{i}'(0)\): if \(D_{a}<D_{a,\max}\);
• from transition to a new \(D^{2}\ell\) by the counter to loading \(D_{a}(0)\).

$k = 0;\ 1;\ 2\ldots.$ In this calculation is carried out for two values of the induction in the air gap, $B_\delta = 1.0$ and $1.2$ T, and, correspondingly, two values of the slot-opening coefficient, $k_p = 0.55$ and $0.47$. The saturation coefficient $k_{\text{sat}} = 0.75$ and the ratio of the wedge height to the slot width, $k_z = 0.5$, remained constant.

The results of the calculation for some dependences are presented as curves in Figs. 3, 4, 5, and 6. For convenience in comparing variants, the power of the impulse generator is expressed in conventional relative units. In each case, the unit is taken to be the power of a generator which, for the given active-part volume, has a definite value of all the parameters affecting the power.

Choice of the Optimal Value of $\beta$

In turbogenerators, the range of possible variation of the value of $\beta$ is determined by the mechanical strength of the rotor. The maximum value $\beta_{\max}$ is determined by the maximum permissible value of the rotor body diameter. The minimum value $\beta_{\min}$ is determined by the maximum permissible length of the rotor body. At present, from the conditions of mechanical strength, the limiting rotor diameter of a two-pole turbogenerator is determined by the value $D_p = 1.15$ m, and the limiting length of the rotor body $l_p$ by the ratio for which

$$ \frac{l_p}{D_p} \leq 7. $$

These limitations on diameter and length will also be valid for impulse generators. The range of possible variations of the value of $\beta$ is shown in Fig. 2. As is known, in designing turbogenerators the range of recommended values of $\beta$ lies within $0.23 \div 0.52$ [1]. The value $\beta = 0.52$ is optimal from the point of view of the weight of copper. Smaller values of $\beta$ correspond to generators of higher powers, when $D^2 l$ is large. In Fig. 2 (from the data of Table 2), the symbol $(x)$ represents

Fig. 2. Range of possible values of $\beta$ for two-pole synchronous machines.

Table 2

The values of \(\beta\) of existing domestic and foreign (Metropolitan Vickers) turbogenerators and generators of breaking capacity are given. For impulse generators that it is expedient to build within the dimensions of high-power turbogenerators, the value of \(\beta\) is practically determined by narrow limits

\[ \beta = 0.23 \div 0.3. \]

In the case of designing an impulse generator for a small power, the range of possible values of \(\beta\) widens; however, the maximum power within the given dimensions will be attained at smaller values of \(\beta\).

Fig. 3. Dependence \(P_{ud}=f(\beta)\) for \(k_1=2\), \(D^2l=3.5\); \(B_\delta=1.2\); \(k_d=0.5\) for different numbers \(z'_1\).

Figure 3 presents a series of curves \(P_{ud}=f(\beta)\) for different numbers of stator slots; here the power of a generator having

\[ D^2l = 3.5\ \text{m}^3;\quad \beta = 0.3;\quad k_d = 0.5; \]

\[ z'_1 = 36;\quad k_1 = 2;\quad B_\delta = 1.2\ \text{T} \]

is taken as unity.

It follows from the curves of Fig. 3 that, for the given volume of the active part of the machine, the optimum value of \(\beta\) will be that at which the generator has the greatest stator length. This is explained by the reduction of the total leakage permeance due to the reduction of the leakage permeance of the end parts with decreasing \(\beta\). The curves of Fig. 4, constructed for different damping coefficients of the leakage fluxes of the end parts (\(k_d=0.5;\ 0.75\) and \(1\)), testify to the same effect; for smaller values of \(k_d\), that is, for smaller permeances of the end leakage...

the scattering curves are located higher and have a smaller inclination to the abscissa axis. When the value of \(\beta\) increases from 0.23 to 0.3, the power of the impulse generator decreases only slightly, on average by 5%. Therefore one may recommend for impulse generators the following limits for the optimum values of \(\beta\):

Fig. 4. Dependence \(P_{\mathrm{ud}} = f(\beta)\) for \(k_1 = 2\);
\(D^2 l = 3.5\); \(B_\delta = 1.2\); \(z'_1 = 36\) for different \(k_d\).

\[ 0.23 \leqslant \beta \leqslant 0.3. \]

Choice of the Optimum Slot Geometry

From the expression for the slot-leakage permeance it follows that the magnitude of the permeance depends chiefly on the ratio \(k_1\). In order to obtain the minimum leakage reactance of the slot, the slot height should be reduced while increasing its width. However, the slot width has a limit, which is determined by the tooth pitch of the stator, i.e., by the number of teeth for a given diameter and by the saturation of the tooth steel. For a given bore diameter and number of stator teeth, it is advisable to choose the greatest possible slot width. With the selected slot width, its depth is determined from two contradictory conditions. On the one hand, increasing the slot height leads to an increase in the volume of copper in the slot and thereby raises the generator power. On the other hand, as the slot height increases, the leakage reactance increases, which leads to a reduction in impulse power.

Calculations have shown that the optimum value of \(k_1\) is not constant, but varies both with the change in the number of stator slots for a given active-part volume and with the change in the active-part volume. In Fig. 5 the curves \(P_{\mathrm{ud}} = f(k_1)\) are presented, constructed for different numbers of stator slots; here the power of a generator having \(D^2 l = 7.0\ \text{m}^3\), \(z'_1 = 36\), \(\beta = 0.3\), \(B_\delta = 1.2\), \(k_d = 0.5\) has been taken as unity. It follows from the curves of Fig. 5 that, for a specified machine volume, each number of slots corresponds to its own value of \(k_1\), varying within fairly wide limits; moreover, the optimum value of \(k_1\) increases with

Fig. 5. Dependence \(P_{\mathrm{ud}} = f(k_1)\) for
\(\beta = 0.3;\ D^2 l = 7.0;\ B_\delta = 1.2;\ k_d = 0.5\) for
various \(z'_1\).

Fig. 6. Dependence of the optimal values \(k_1 = f(D^2 l)\) for
\(\beta = 0.3;\ B_\delta = 1.2;\ k_d = 0.5\) for
various \(z'_1\).

with an increase in the number of stator slots. For example, for \(z'_1 = 24\) and 30 the optimal value is \(k_1 = 1\), while for \(z'_1 = 54\) and 60, \(k_1 = 1.5 \div 1.8\). The dependences \(P_{\mathrm{ud}} = f(k_1)\) have an analogous form for all the other volumes considered.

Figure 6 presents curves characterizing the dependence of the optimal values of \(k_1\) on the quantity \(D^2 l\) for various numbers of stator slots. These curves were constructed from the data following from Fig. 5 and figures analogous to it, constructed for other volumes. From the curves in Fig. 6 it follows that, when the value of \(D^2 l\) varies within the limits \(2 \div 11\ \mathrm{m}^3\), the optimal value of \(k_1\) for the considered series of numbers of stator slots changes within the limits \(0.9 \div 2.5\).

For machines with the same number of stator slots, the optimal value of \(k_1\) decreases as the volume increases. As the number of stator slots increases, the limits of the optimal values of \(k_1\), depending on the volume of the active part of the machines, become broader. For example, when the value \(D^2 l\) varies within the limits \(2 \div 11\ \mathrm{m}^3\), for \(z_1 = 24\) or 30 the optimal value of \(k_1\) is close to unity, whereas for \(z_1 = 54\) or 60 the range of values of \(k_1\) has the limits \(1.2 \div 2.5\).

Choice of the optimal number of stator slots

For the selected values of \(\beta\) and \(k_1\), one can find such a number of stator slots at which the power of the impulse generator will be maximal.

If one constructs curves analogous to those presented in Figs. 3, 5, and 6 for all the considered values of the quantities that remain constant in each separate figure, then it is possible to determine, for any value of \(D^2 l\), the optimal number of stator slots. The calculations carried out showed that, for impulse generators in which the quantity \(D^2 l\), proportional to the volume of the active part, lies within the limits \(2 \leq D^2 l \leq 9\ \mathrm{m}^3\), the optimal number of stator slots is determined by the value \(z_1 = 48\) and 54. For generators in which the quantity \(D^2 l \geq 9\ \mathrm{m}^3\), the optimal number of slots is \(z'_1 = 54\) and 60.

In conclusion, it should be noted that, in designing impulse generators, as in designing machines of normal construction, deviations from the optimal ratios are possible. However, in most cases the optimal values found for the principal quantities determining the geometry of the machine will serve as initial data in determining the principal dimensions of generators of impulse power and may prove useful in calculations of generators of burst power.

References

1. E. G. Komar. Problems in the Design of Turbogenerators, GEI, 1955.
2. I. M. Postnikov. Proceedings of LPI, No. 2, 1937, No. 209, 1960.
3. I. M. Postnikov. Design of Electrical Machines. GITL USSR, 1960.
4. L. G. Mamikonyants. Elektrichestvo, No. 2, 1956; No. 7, 1957.
5. G. A. Sipailov, K. A. Khor’kov. “Power, Energy, and Machine Constant of an Impulse Generator,” Izvestiya TPI, 1964.